Long-term effects of shade and input levels on coffee yields in the Pacific region of Nicaragua

The suitability and profitability of coffee cultivation in Central America are at risk due to pest and disease outbreaks, price fluctuations and climate change. Proper shading is claimed to be one of the most promising practices to seek sustainability and better adapt coffee cultivation to climate change in marginal areas. This study recorded and compared coffee cherry yields over a ten-year period from shaded coffee (N-fixing-trees and timber trees) agroforestry systems under different management regimes (conventional vs. organic) in a suboptimal site. Significant differences in production were detected between conventional inputs vs. combination of organic inputs and shade types in some years of the evaluation period. Full-sun cultivation under intensive management was the most productive system for coffee yields, followed by shaded systems under timber trees. Interestingly, and regardless of management systems (intensive conventional or intensive organic) the worst combinations in terms of coffee yield were shaded systems under leguminous species (Inga laurina (Sw.) Willd. + Simarouba glauca DC.). Across all experimental plots, the timber species Simarouba glauca and Tabebuia rosea (Bertol.) DC. grew well, reaching a mean annual increment in diameter of 2.5-3.3 cm/year (age 12 years). Average gross revenues were higher in full-sun and timber-shaded agroforestry systems. Overall, intensive management regimes were the most expensive cultivation system to run but also the best in terms of coffee yield performance.


Long-term effects of shade and input levels on coffee yields in the Pacific region of Nicaragua
Bois et Forêts des Tropiques -ISSN: L-0006-579X Volume 346 -4 th quarter -December 2020 -p. 21-33

Long-term effects of shade and input levels on coffee yields in the Pacific region of Nicaragua
The suitability and profitability of coffee cultivation in Central America are at risk due to pest and disease outbreaks, price fluctuations and climate change. Proper shading is claimed to be one of the most promising practices to seek sustainability and better adapt coffee cultivation to climate change in marginal areas. This study recorded and compared coffee cherry yields over a ten-year period from shaded coffee (N-fixing-trees and timber trees) agroforestry systems under different management regimes (conventional vs. organic) in a suboptimal site. Significant differences in production were detected between conventional inputs vs. combination of organic inputs and shade types in some years of the evaluation period. Full-sun cultivation under intensive management was the most productive system for coffee yields, followed by shaded systems under timber trees. Interestingly, and regardless of management systems (intensive conventional or intensive organic) the worst combinations in terms of coffee yield were shaded systems under leguminous species (Inga laurina (Sw.) Willd. + Simarouba glauca DC.). Across all experimental plots, the timber species Simarouba glauca and Tabebuia rosea (Bertol.) DC. grew well, reaching a mean annual increment in diameter of 2.5-3.3 cm/year (age 12 years). Average gross revenues were higher in full-sun and timber-shaded agroforestry systems. Overall, intensive management regimes were the most expensive cultivation system to run but also the best in terms of coffee yield performance.

Introduction
Climate change will negatively impact both the yield and quality of arabica coffee production in most of the growing regions world-wide (Vaast et al., 2005;Ovalle-Rivera et al., 2015). Climatic variability will also trigger pests and disease outbreaks and alter key ecological interactions within shaded coffee agroecosystems (Avelino et al., 2015;Magrach and Ghazoul, 2015). In Nicaragua, it is projected that the climate will become hotter, dryer and with remarkable seasonality especially at lower altitudes which in turn will move upward the altitudinal range of suitable land to grow coffee (Laderach et al., 2011). This scenario becomes worse for lowland areas (400-500 m above sea level -asl-in the Pacific region) where coffee is still grown (Padovan et al., 2015). Nicaragua does not have available land at an altitude higher than 1,300 m asl to move coffee cultivation upwards, therefore designing and managing diversified and more resilient shaded coffee agroforestry systems to sustain farmers' livelihoods in the long term becomes a crucial action in marginal areas (Bouroncle et al., 2017;Läderach et al., 2017).
Across Mesoamerica, frequently used shade tree species for both coffee and cacao cultivation will also be affected by climate change (de Sousa et al., 2017(de Sousa et al., , 2019. For example, according to recent modelling of the suitability of common shade tree species, it is expected that the range of occurrence of 79% of the common shade tree species used for coffee, will grow in a narrower geographical belt and thus will be lest suitable to intercrop with coffee (Gay et al., 2006;de Sousa et al., 2019). The biggest losses in shade tree adaptability will be experienced in mid-altitudinal coffee-growing areas (400-700 m), especially by the most popular N-fixing, fruit, and timber shade trees. Climate change is already having significant adverse impacts on smallholder coffee and basic grain farmers across Central American, yet mitigation and adaptation measures are gradually taking place (Baca et al., 2014;Harvey et al., 2017).
Globally, coffee is cultivated under five main production systems defined by their vegetation composition, structural complexity, and management level. These coffee-production systems cover the spectrum from a very diverse "rustic" shaded-system to intensive monoculture either under a monospecific shade tree or in full-sun (see Moguel and Toledo, 1999 for an in-depth description of the systems). The extremes of this spectrum also represent the traditional management regimes with high diversity, which produce low coffee yields, and monocultures characterized by high coffee yields.
In Central America, coffee is produced mainly under the shade of trees (Jha et al., 2014;Somarriba and López-Sampson, 2018). The contribution of companion trees has been largely recognized on soil health (Moço et al., 2010;Thomazini et al., 2015), coffee quality (Vaast et al., 2005;Bosselmann et al., 2009), and these trees support the livelihoods of coffee smallholders (Méndez et al., 2009(Méndez et al., , 2010 while providing a buffer to counterbalance harsh conditions caused by climate variability in the long term (Mbow et al., 2014). Proper management of companion trees can improve the growing conditions of the multi strata system reducing abiotic stress and facilitating the performance of understory crops (Beer et al., 1998;Tscharntke et al., 2011). Besides, farmers can benefit from agroforestry by its capacity to provide several ecological services, such as water conservation, nutrient cycling, carbon sequestration, pollination and regulation of pests and diseases (Jose, 2009;Vaast and Somarriba, 2014). Most management regimes for coffee cultivation are mainly focused on increasing productivity and incomes with less attention given to crop resilience and sustainability. Jezeer et al. (2018), however, found similar economic performance between shade (low, medium and high) and input (high, medium and low) classes, but a reduced net income and benefit-cost ratio in the high-input class, confirming a win-win scenario of economic performance with more environmental-friendly production through agroforestry. Moreover, net income from organic coffee production systems was similar to conventional production systems (when excluding the cost of certification), even when the former had a 22% lower yield than their counterparts, mainly due to the premium price paid to organic farmers which compensated the lower yields (Lyngbaek et al., 2001). Similarly, Rossi et al. (2011) evaluated the effect of management practices on coffee productivity in a long-term experiment in sub-optimal growing conditions and found that intermediate management intensity produces competitive coffee yields overtime.
In this research, we evaluated the performance of coffee yields over 10 years under different management regimes and shade canopy trees (N-fixing trees and timber trees) to determine the best possible shade tree combinations for coffee productivity. The following questions were addressed in this study: (1) Are there significant effects of shade composition and management regimes on coffee productivity? (2) In the timber-based agroforestry system (AFs), do the potential revenues from standing volume yields lead to higher total system revenues (sum of all marketable goods) compared to leguminous-based AFs?

Study area and site description
This study was carried out from November 2001 to May 2014 in three experiment plots located at two different nearby sites in Masaya and Carazo, southern Nicaragua. The first two plots were established in a local Botanical Garden (11o 53' 54" N, 86 o08' 56" W) managed by the Tropical Agricultural Research and Higher Education Center (CATIE in Spanish), in partnership with the Agrarian National University (UNA), Federation of Cooperatives of Credit and Savings (CENECOOP-FEDECARUNA) and Nicaraguan Institute of Agrarian Technology (INTA). The third site was established in 2001 in the Campos Azules Experimental Center (CECA) handled by the INTA. The three sites are located at 455 m asl which is rather marginal for Arabica coffee cultivation. This region has a long history of coffee cultivation as the Central-Pacific Region of Nicaragua was the first region where coffee was introduced in late 1880 (Craipeau, 1992). Despite being classified as marginal land to grow coffee several farmers still rely on coffee cultivation as living income.
The experiment covers 3 ha. Details of experiment establishment were described by Haggar et al. (2011). The mean annual temperature was 27°C and mean annual rainfall of 1,470 mm (Haggar et al., 2011). Most of the total annual precipitation falls over the wet season (May-November) while a pronounced seasonal drought occurs from late November to mid-May. Predominant soils in the study sites are originated from volcanic eruptions and classified as Andisols. These soils are commonly deep, well-drained, and have high organic matter content, low bulk density, high allophane content, and consequently a high phosphorus fixation capacity, high amorphous mineral content, and high-water retention capacity (Padovan et al., 2015).

Experimental design
The five different coffee production systems under comparison included four agroforestry combinations (AFs) and a full sun plot as control, both under conventional (CONV) and organic (ORG) management (tables I and II). Fourteen combinations of shade trees, input levels, and management regimes were implemented. The combinations evaluated are described in table I. Three replicates were set at each site as a randomized block design with shade type as the main treatment and input levels as sub-treatments within shade types. Four different species were planted, managed, and combined with two different management regimes namely CONV and ORG, and input levels (moderate and intensive) (table II). The combination of tree species was selected depending on their morphological complementarity and their ability to provide, or not, certain products and services such as timber, firewood, or nitrogen fixation. Selected tree species were generally used by local coffee farmers in nearby localities (Bonilla and Somarriba, 2000;Cordero et al., 2003).
Subplots sizes varied between 500 to 600 m 2 with an effective measurement area ranging from 225 to 300 m 2 and comprising a minimum of 24 shade trees and 100 coffee plants per plot. The final number of plots and treatments included in the experiment was limited by land availability and labor costs. Approximately 1,667 shade trees were initially planted across experimental plots, however, three thinning events were done at age 5, 8, and 12 years after planting to keep a final density of 272 shade trees/ha. Coffee plant density (Coffea arabica var Pacas) was 4,000 plants/ ha. In this study, we only showed the results of coffee yields until the harvest season of 2011 given that in 2012 there was a drastic drop in coffee yields across all management regimes and shade combinations. Shade tree growing performance and management data were included until the year 2012 of evaluation.

Data collection
During the period 2002-2012, in each plot, coffee yield from 100 coffee plants was recorded annually. Coffee was manually harvested, depulped, fermented, washed and sun-dried. Yield per plot was then converted to kg/ha/year. As a part of regular maintenance, coffee plants were partially pruned every year, thus pruning intensity was registered from 2005 to 2012 as a proportion of planting density and accounted for coffee yield recordings. Tree dimensions and shade cover within plots were measured annually during the first year and then every two years. Shade cover (%) was measured at the center of each plot with a densiometer and taking four readings in each cardinal point. Shade cover from each measuring point was averaged and registered as the plot's annual value. Tree growth of each shade species (diameter in cm, total and commercial height both in m and crown area in m 2 ) was recorded every 6 months during the first five years and then annually until they reached the age of 14 years. Tree diameter and tree height were measured with diameter tape and clinometer, respectively. Between 20-24 shade trees were measured per plot and recorded tree attributes were averaged and coded per species.

Data management and statistical analysis
To evaluate the effect of shade tree species, input levels, and management regimes (grouped in a single variable named Treatment) on coffee yield (t/ha) a general linear mixed model was used (Di Rienzo et al., 2011). In our model specification, the Treatment variable is considered as a random effect to account for the imbalance in the statistical design. The statement of the model was: Coffee yield ~ 1 + Treatment. The analysis was carried out by year due to the year variable was masking the "Treat-ment" effect when running treatment as a fixed effect and year and block as random effects. Fisher's Least Significant Difference (LSD) means comparisons were run among management regimes to evaluate statistical differences.
Annual total costs (coffee maintenance + harvest) were totaled, and gross revenues (incomes from coffee + firewood) was calculated per hectare and treatment. All the labor was hired. Gross revenues represent the earns from the sale of dried coffee and firewood sold in the local market at a particular year. The annual costs represent the actual cost incurred to maintain and run the experimental coffee plantation according to the management regime and input levels. Total Costs and Gross revenues are presented in US Dollars. Annual Increment in diameter (cm/ year) and basal area (m 2 /ha/year) of timber and N-fixing trees are presented by shade type and input level combination. All statistical analyses were performed in InfoStat (Di Rienzo et al., 2011).

Coffee yield, management regimes and shade typology
Coffee yields varied between years, shade typology and management regimes (figures 1 and 2). There was a significant effect of shade combination and management regimes on coffee yield only in three out of ten years evaluated (table III). Fisher's LSD means comparisons among treatments did not show any pattern on yields performance in terms of best shade combinations and management regimes. For example, in 2002 all 14 fourteen combinations registered similar coffee cherry yields. However, in the second, fourth, sixth and seventh harvesting years, the full-sun coffee system under conventional-intensive management had the highest yields and was statically different from the other treatments. Intensive Organic (IO) management under the shade of timber trees combination (i.e. SSTR Samanea saman + Tabebuia rosea; and SGTR   (table IV). Until 2005, all tree shade combinations and management regimes showed an increase in coffee yield, although the intensive conventional (IC) and intensive organic (IO) management regimes showed better yields. However, during the period 2009 to 2011, there was a drop-in yields in all systems and it was in those bad years that coffee production under organic inputs regimes ranked better in terms of yield performance. Yet, lower coffee production was observed under ILSG at its two input levels (MC and IO). The Inga-Simarouba combination negatively influenced production regardless of whether the management regime was organic or conventional. ILSG performed better only when there was a drop-in production in the other treatments as a positive response to particular agronomic practices (i.e. pruning) or because of a stressful event (i.e. an extended dry period in a particular year of the evaluation). Better responses in terms of coffee yield were   T. rosea, and S. saman). In some years, timber-based coffee system (SGTR and SSTR) performed similar to or better than the full-sun coffee system (p > 0.05). It seems that coffee yield is influenced by different levels of input, but when the system includes the shade of I. laurina there was no response on coffee yield along the evaluated period.

Performance of shade trees and silvicultural management
A 40% shade level threshold was intended to be maintained in all shaded coffee systems combinations through thinning and pruning. Three thinning events were performed during the cycle of evaluation, the first one was done in 2005 (two-three years after planting) with an intensity of 45-50% across treatments and management regimes, and the second one was performed in 2008 (5-6 years after planting) with an intensity of 25%. The last thinning event occurred in 2014 (12 years after planting, 25% intensity) which resulted in a final shade tree density across treatments ranging from 17 to 33 trees per plot (table V). However, an increase in the shade level above the attempted threshold was evidenced in 2014 and 2016 due to a lack of management of the companion trees pro-bably because of the drop-in coffee yields. Regardless of the management regime, timber species grew better than leguminous shade tree species (table V). Mean Annual Increment in diameter (MAI-d) of I. laurina varied between 2.9 to 3.2 cm/year, for S. glauca MAI-ranged from 2.5 to 2.7 cm/year, for T. rosea MAI-d values were 2.8 to 3.3 cm/ year and MAI-d values for S. saman ranked last with 1.3 to 1.9 cm/year (table VI).

Coffee production costs and revenues under different management regimes and shade type
The intensive organic system was the most expensive management regime to run per ha and year 1,474.2 USD/ha (± 122.9), with inputs representing most of the costs, irrespective of the shade typology. Followed by the full-sun coffee production system with expenses tallying 1,303.5 USD/ha (± 149.3). Moderate organic and moderate conventional registered the lowest costs across different shade typologies and full-sun systems (879.8 USD/ ha ± 91). In terms of gross revenues, both full-sun systems and timber-based coffee agroforestry systems (SSTR and SGTR) were the treatments with better financial returns, even though gross incomes from the sale of timber trees were not considered (figure 3).

Table V.
Tree population remaining after the three thinning events performed in shaded coffee agroforestry systems in Nicaragua.

Coffee production: Interaction between shade typology and management regime
Full sun and conventional management have been largely promoted as an avenue to intensify coffee cultivation in producing countries (Perfecto et al., 2005;Guhl, 2008). It estimated that 41% of the coffee area worldwide is produced without the shade of trees (Jha et al., 2014). Although coffee intensification leads to an increase in yields, achieved by the high use of external inputs such as fertilizers, pesticides, and fungicides, it also has an increasing environmental footprint (Lyngbaek et al., 2001;DaMatta, 2004). In our study, full sun production under intensive management performed better in most of the years of evaluation. While cof-fee yields of organic systems are usually lower than conventional ones (Lyngbaek et al., 2001;Van Der Vossen, 2005), this study demonstrated that coffee yields under intensive organic management regimes were similar to those obtained under intensive conventional management yet some significant differences were evident overtime (figure 2, table II).
The best coffee yields came from the combination of full-sun and intensive conventional management, followed by timber tress (SSTR and SGTR) under intensive organic systems. This demonstrates that the use of high-level inputs and moderate shade had a strong influence on coffee yields. Contrary to expectations, the ILSG combination (leguminous shade tree) had the worst coffee yield performance, mainly due to negative effect of I. laurina on coffee plants, as this species had a dense canopy thus increasing the shade level of the plot. Haggar et al. (2011) also pointed out that the shade cast by I. laurina prevented a better response of coffee production even when conventional inputs and intensive organic is provided. Similarly, Siles et al. (2009) found lower coffee yield under agroforestry systems due to the lack of pruning and thinning of the companion trees. Therefore, there is a need to efficiently manage (frequency and intensity of events) shade cover according to the characteristics of the companion species (size of the trees, crown shade and canopy density) (Somarriba and Calvo, 1998;Somarriba and Beer, 2010) as there is a direct negative effect between the increase in planting density of shade trees, shade cover (%) and expected coffee production (Jezeer et al., 2018).
The advantage of shading in coffee yields is only realized if a judicious evaluation and management plan of the companion trees is done and matched with local agro-ecological characteristics when designing coffee agroforestry systems (Beer et al., 1998;DaMatta, 2004;Bosselmann et al.,  2009). In general terms, the benefit of tree cover increases when the agro-environment is less favourable for sustainable coffee production (DaMatta, 2004;Carr and Lockwood, 2011). For example, shade provided by trees acts as a buffer for sub-optimal environments such as extreme temperatures and prolonged droughts (Siles et al., 2009;Padovan et al., 2015). In addition, van Kanten and Vaast (2006) reported that for coffee grown in suboptimal low-altitude sites, Euca-lyptus deglupta cast the most beneficial shade as it maintained a more constant shade level throughout the year thus better protecting coffee plants in comparison to Terminalia ivorensis and Erythrina poeppigiana which underwent complete defoliation during the dry period. In our study, it seems that the type and combination of trees influenced coffee yield performance. Cordero et al. (2003)   widely used as a shade to protect coffee plants during the 6 months-long dry seasons in the south Pacific of Nicaragua. Padovan et al. (2015) studied root distribution and water use in the same experimental plot and found a root niche differentiation for optimal exploitation of resources in shaded coffee agroforestry systems. However, there were differences between shade tree species, S. glauca exhibited a denser root system in deep soil layers compared to T. rosea, indicating complementarity with coffee root growth. The shaded coffee system used more water and at a greater rate at the beginning of the two dry seasons evaluated, explained by a better exploration of the roots of shade trees in the soil profile. Yet, when the shaded coffee system faced a severe dry season the advantage of better exploration was inexistent as it seems that underground water from the systems was no longer available for the crops and the shade trees (Motisi et al., 2019).

Costs and gross revenues under different shade typologies and management regimes
In this study, the costs of running intensive organic and intensive conventional management systems for coffee production had similar figures with the highest costs represented by inputs (fertilizers, pest, and diseases control, weeding) and exceeded the conventional moderate and moderate organic management systems in 28%-37% and 33%-40% of the costs, respectively. Other studies reported coffee yield under conventional management exceeding in more than 20% of the costs spent in comparison with organic management regimes (Lyngbaek et al., 2001). Likewise, Jezeer et al. (2018) found for conventional coffee cultivation, higher yields associated with higher costs, mainly due chemical inputs and hired labour. In our study higher yields were obtained in both intensive management systems, however large variation in coffee yields was recorded between years (table IV). Full-sun systems and timber-based coffee agroforestry systems had the best average gross revenues as compared to leguminous based-coffee systems.
Timber trees are a conspicuous element in shaded coffee agroforestry systems (López et al., 2003;Peeters et al., 2003;Méndez et al., 2009). In our study of coffee yield performance under timber, trees were similar to that of full-sun cultivation systems. Regardless of the management regime, timber species grew better than leguminous shade tree species (table VI). Timber tree growth rates in shaded coffee systems were in the range of tree growth reported for other timber trees in multistrata-agroforestry systems elsewhere (López et al., 2003;López-Sánchez and Musálem, 2007;Méndez et al., 2009). Even in our analysis, we did not consider the potential income from harvesting timber trees in the gross revenue estimates, selling timber could add on average 400 USD/ha (std 307 USD/ha) to the overall gross income of timber-based coffee systems. Several authors support the argument that both firewood and timber generate considerable revenues for coffee growers which helps cope with fluctuation of coffee prices and pest and diseases outbreaks (Peeters et al., 2003;Rice, 2008;Noponen et al., 2013;Ehrenbergerová et al., 2018Ehrenbergerová et al., , 2019.

Conclusion
Coffee productivity was influenced by levels of inputs and management systems (conventional and organic) and shade typologies. Overall, in this study, intensive organic production was found to be equally productive as conventional production regime. Intensive full sun-conventional coffee systems had the highest productivity, followed by the combination of timber Samanea saman + Tabebuia rosea organic intensive, which showed the positive effect of shade typology and intensive organic inputs on coffee yields. Costs of running conventional inputs vs intensive organic inputs were similar, however the use of intensive organic inputs lowers the net profitability of the systems due to their high maintenance costs. Moderate organic treatments showed a better performance when combined with timber trees as they could provide an increase in the total income by selling standing timber stock. The economic potential value of Tabebuia rosea and Simarouba glauca, in the long term is around 800 USD/ha and 400 USD/ha, respectively. This is a clear indication that coffee cultivation under shade helps to maximize the benefits of diversification while minimizing competition between companion trees and coffee plants.